Lloyd, Bekenstein and Landauer agree

Occasionally three arguments with completely different machinery land on the same number, and when they do, the number stops being a guess. This post is about one such convergence, and it is the load-bearing fact behind the claim that a horizon has a computational budget.

The three arguments

First, Lloyd’s bound. A system with energy E, evolving for time t, can perform at most 2 E t over pi hbar operations. This comes from Margolus and Levitin, and it is quantum kinematics: the maximum rate at which orthogonal states can be cycled is set by the energy spread, the same mathematics that sets the speed of quantum gates. Apply it to the energy inside a cosmological horizon, E is the horizon mass times c squared, and t is one Hubble time.

Second, the Bekenstein bound. The maximum information in a region is its horizon area in Planck units, divided by 4 and by the logarithm of 2 to convert nats to bits. This is black hole thermodynamics and the holographic principle: the information content of a region is its boundary area, not its volume.

Third, Landauer’s principle. Erasing one bit costs k T ln 2 of energy. If you ask how many bit-erases the horizon’s free energy budget supports over a Hubble time, at the Gibbons-Hawking temperature that the previous post derived, you get a number of bit operations. This argument knows nothing about quantum gates or holography. It knows about heat and cost.

The number

Run all three. The de Sitter horizon entropy is S d S equals pi times c to the 5, over G h bar H squared, which for the measured Hubble constant comes to about 2.27 times 10 to the 122 nats.

The Bekenstein-Lloyd side converts directly: S over ln 2, which is 3.27 times 10 to the 122 bits.

The Lloyd side: 2 E t over pi hbar, with E the horizon energy, works out to S over pi squared, which is 2.30 times 10 to the 121 operations per Hubble time.

The Landauer side, taking the horizon’s free energy at the Gibbons-Hawking temperature over one Hubble time, also lands on the S over pi squared value, and the coincidence has a reason: the Gibbons-Hawking temperature is defined by the same horizon geometry that sets S, so the two budgets are the same budget counted in different units. The ratio of operations to bits is ln 2 over pi squared, about 0.07, for every horizon: about one operation per 14 bits per Hubble time.

The agreement is not three miracles. It is one geometric fact, the horizon entropy, dressed in three different outfits. But the outfits are genuinely independent constructions, from gate kinematics, from thermodynamic area, from erasure cost, and the fact that all three can be worn by the same number is what makes the number respectable.

What this buys the programme

A computational multiverse claims that branches are computed. Any such claim owes a budget: how much computation does the substrate have available. The answer above is it. 2.3 times 10 to the 121 operations and 3.3 times 10 to the 122 bits per horizon per Hubble time. Two consequences, developed in the computational series of this archive. Any candidate substrate numerology that uses a number far below this budget, and the framework’s retracted one did, is leaving most of the capacity unexplained, which is a form of arbitrariness. And any rendering scheme whose demands exceed this budget cannot be faithful, which is how the fact-budget post later caps what a simulation can honestly render.

The number has a standing weakness worth stating: it is a budget for one horizon, not a derivation that our horizon is computed. Bounds of the form at most say nothing about whether. What they say is that whatever the answer is, it fits in these margins.